Changing the shape of the commercial cycle

"Those of us still in the macro-economic forecasting game had better be careful about our choice of indicators"

The labour market has been behaving very strangely of late, raising all sorts of questions for those of us who study the ups and downs of economic and electoral cycles. This behaviour provides clear evidence of the more flexible jobs market for which the Government has been striving. But what will the electorate make of all this?

My old text-books tell me labour is a variable factor of production, but of course in practice employers usually find it costly to vary employment, particularly in a downward direction. So when demand first turns down, managers typically cut overtime or put staff on short-time working. Productivity falls as numbers employed lag output. Capital expenditure plans are usually protected, too, and investment often continues to rise long after the economy slides into recession.

But short-time working is expensive so, as the pattern of recession is established, working hours stabilise and employment and investment levels fall. Vacancies and unemployment figures, together with investment, usually trail well behind output. The Central Statistical Office, in its regular analysis of the UK business cycle, has always used unemployment and investment as lagging indicators of the cycle. Until the system was revised in 1980, vacancies were also part of its lagging indicator index.

When demand eventually turns up, the first response of employers is naturally a tentative one. Rather than looking for new staff or new equipment, they usually ask the workforce to clock up more overtime. Working hours and take-home pay usually bounce back sharply at this stage of the cycle, providing a clear signal to both economic analysts and shopfloor workers that the economy is on the mend.

You might think it appropriate at this stage to reach for a more modern economics book. However, the Government has stuck to the old texts and refused to subscribe to any new theories. And the recent behaviour of the labour market suggests that, with employee protection laws substantially eroded, companies have begun to regard labour as a flexible factor of production. Those old textbooks could come in handy after all.

The sequence of events in the jobs market has certainly been very different this time. Incredibly, vacancies now act as a leading rather than a lagging indicator of the cycle. At the same time, the lag between unemployment and output seems to have shortened dramatically. On the other hand, working hours, traditionally a sensitive barometer of the turning point, have not responded to the recent current change in economic pressure.

Vacancies began to signal a downturn in 1988, and had fallen back by well over a third by the time output finally turned down in mid-1990. Employment levels then moved down alongside output, with hardly any movement in productivity. Vacancies began to revive in mid-1991, well before the output recovery began in 1992. Companies anticipated demand on this occasion by trying to recruit more labour rather than by boosting overtime.

This helps to explain why unemployment, which trailed GDP by five years after the end of the last recession, followed it after just 12 months this time.

These changes are nowhere more apparent than in manufacturing industry. Post-war employment trends in this sector have exhibited a step-wise pattern: sideways movements during the expansion phase have been followed by near- vertical falls during the slow-down phase. The length of the working week has also been falling, but has exhibited a pronounced cyclical pattern, particularly at turning points. This time, employment levels have shown an increase, albeit small, rising by 29,000 during 1994. But here again, the real surprise has been overtime working, which has simply refused to rally in the usual way. As the charts show, this is the dog that did not bark.

The consequences of these developments are likely to be very wide-ranging, but are hard to predict, even in simple economic terms. By making it easier to sack people, the Government surely exaggerated the recession, forcing the burden of adjustment away from companies, onto individuals. Yet by making it less risky for employers to take on new staff these reforms could actually boost employment. The most recent job trends lend support to that theory. But are these new jobs worth having? Ultimately that is for the electorate to decide.

It is hard to think that a revival in the labour market could vindicate its economic policies and save the day for the Government. Millions have been marginalised, losing the little economic power they once had as company insiders.

Indeed, we are all suffering from the job insecurity caused by these policies, despite the recovery trumpeted by the Government. And the latest vacancies figures tell us that the economy and the labour market are now slowing.

Yet it is possible that key blue-collar voters have misread the situation, and that the elusive feel-good factor could return even if the economy does slow.

A research paper that John Curtice and I produced for Kleinwort Benson last year showed that survey-based job security and consumer confidence were strongly correlated with overtime working. Unemployment and other headline data had no significant impact.

It seems that when workers are asked to work longer hours they conclude that the economy is picking up and they are not about to face the sack.

Such evidence is probably far more convincing than reading about falling unemployment totals. And in its absence it would seem that workers remain as insecure about their jobs as they were during the recession.

Will this remain the case? Won't our statistical model break down like so many others before? After all, if employers are varying jobs rather than hours, then the vacancies board outside the factory, empty for so long, should become a much better gauge of job security and the state of the economy than overtime working.

If they see their own employer advertising new jobs, people will surely conclude that their own jobs are safe, even if overtime hours do not pick up.

Meanwhile, those of us still in the macroeconomic forecasting game had clearly better be careful about our choice of indicators.

The economic structure is not set in stone. Conventional fixed-lag econometric models are very hard to use when this is shifting, and even the more flexible leading indicator and turning-point recognition techniques imported from the United States are suspect in this situation.

As Nigel Lawson found to his cost when liberalising the financial sector, governments may not be able to outlaw the business cycle, but they can certainly change its shape.

The author is Professor of Financial Economics at Birkbeck College.